1 / 13

Linear prediction

Linear prediction. Linear prediction is a method used to reduce the bandwidth required to transmit PCM pulses It is widely used in speech communications over mobile channels In linear prediction the future values of a discrete time signals are estimated as a linear function of previous samples.

Télécharger la présentation

Linear prediction

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Linear prediction Linear prediction is a method used to reduce the bandwidth required to transmit PCM pulses It is widely used in speech communications over mobile channels In linear prediction the future values of a discrete time signals are estimated as a linear function of previous samples

  2. Linear prediction filter The predicted future samples of the discrete signal can be predicted from the present samples of the discrete signals by passing into the LP filter shown below

  3. Linear prediction filter The equation that describes the process is given by (convolution ) Where p, is the number of unit delay elements which is called the prediction order w represents the weights of the taps

  4. Linear prediction error The prediction error is defined as The question that may be asked is how many filter taps are required? The design objective is to choose the filter coefficients w1-wp so as to minimize the mean square error

  5. Linear prediction error If we expand the previous equation for J If the input signal x(t) is the sample function of a stationary process X(t) of zero mean, that is, , then

  6. Linear prediction error • The term is the autocorrelation function of x[n] • The mean square error J can be rewritten as • In order to find the coefficients which minimizes the error, we differentiate J with respect to wk and equate with zero

  7. Wiener-Hopf equation If we apply the derivative to the mean square error J we get the following equation The above optimality equation is called the Wiener-Hopf equation for linear prediction It is convenient to reformulate the Wiener-Hopf equations in matrix form as follows

  8. Matrix form of Wiener-Hopf equation Where w0 is p-by-1 optimum coefficient vector rx is p-by-1 autocorrelation vector

  9. Matrix form of Wiener-Hopf equation Rx is p-by-p autocorrelation matrix

  10. Wiener-Hopf equation and the LP filter design In order to design the filter we need to compute the tap weights which can be determined from The minimum mean-square value of the prediction error is obtained by substituting the values wo into J which yields the following equation

  11. example

  12. Solution

  13. Solution

More Related